On Mon, Mar 23, 2015 at 12:02:05PM -0700, Vagrant Cascadian wrote: > >> On Tue, Mar 17, 2015 at 10:00:08PM +0100, Wolfgang Schweer wrote: > >> > With this (trivial) modification installation succeeded. > >> > > >> > --- a/postinst 2015-03-06 10:00:00.000000000 +0100 > >> > +++ ltsp-client-builder.postinst 2015-03-17 21:39:38.824682599 +0100 > >> > @@ -64,7 +64,7 @@ > >> > db_progress STEP 1 > >> > > >> > if [ "$USE_CDROM" != "false" ] && [ ! -f /target/media/cdrom/.disk/info > >> > ]; then > >> > - chroot /target mount /media/cdrom > >> > + mount /dev/cdrom /target/media/cdrom > >> > log "mounting /media/cdrom" > >> > fi > > This patch breaks USB stick installs. Confirmed after having started an USB stick installation on real (and very old) hardware; something like /dev/sdXY is mounted on /cdrom inside d-i environment.
Inspired by this observation I tried this patch in the virtualbox environment (booting from 'cd') and as well on bare metal w/ USB stick: --- a/ltsp-client-builder.postinst 2015-02-07 18:30:46.000000000 +0100 +++ b/ltsp-client-builder.postinst 2015-03-24 16:04:09.510087326 +0100 @@ -64,7 +64,7 @@ db_progress STEP 1 if [ "$USE_CDROM" != "false" ] && [ ! -f /target/media/cdrom/.disk/info ]; then - chroot /target mount /media/cdrom + mount $(mount | grep cdrom | cut -d' ' -f1 | head -1) /target/media/cdrom log "mounting /media/cdrom" fi @@ -244,3 +244,4 @@ #DEBHELPER# exit 0 + Installation succeded in both cases. 'head -1' is most probably unneeded, added just in case the device is mounted twice. It won't do any harm though, I guess. Wolfgang
signature.asc
Description: Digital signature